chrome cannot set session id cookie correctly when two web applications use same domain
Reported by
garychen...@gmail.com,
Mar 14 2017
|
|||||
Issue descriptionUserAgent: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/56.0.2924.87 Safari/537.36 Steps to reproduce the problem: We have two web applications use tomcat6 and they use the same domain(xxx.test.com),the first application visit by https,the other application visit by http.and they use the same sessionCookieName:JSESSIONID,but the path are not same(https:/ and http:/admin). when the first application visit successfully by https, chrome can set sessionid(JSESSIONID=xxxx),and now don't close the chrome,visit the second application by http,chrome cannot set sessionid with name JSESSIONID.but if we visit the second application by http first,two sessionid will set successfully. 1. clear the cookies 2. visit first web application "https://xxx.test.com" you will find chrome can set sessionid(name:JSESSIONID, path:/) and don't close the chrome 3. visit second web application "http://xxx.test.com",you will find chrome cannot set sessionid(name:JSESSIONID, path:/admin) successfully. 4.if we visit the second web application first, two sessionid with the same name both can set successfully. What is the expected behavior? two sessionid can set successfully. What went wrong? we found chrome can work normal with previous version. and firefox,Safari can also work normal Did this work before? N/A Chrome version: 56.0.2924.87 Channel: n/a OS Version: 6.1 (Windows 7, Windows Server 2008 R2) Flash Version: please check this issue.thanks
,
Mar 14 2017
,
Mar 15 2017
this is the net-internals log, thanks
,
Mar 16 2017
,
Apr 17 2017
garychenqin: Sorry, this dropped off my radar. It looks like the HTTPS URL is setting the JSESSIONID cookie with a secure attribute, and then the HTTP URL is trying to overwrite it with an insecure cookie, something we currently don't allow. Either they should be using different cookie names, or it should not be secure. [+mkwst]: We've run into a bunch of issues with this behavior change. Should we make the issue more visible, or rethink the behavior?
,
May 2 2018
Late to respond, but this is expected behavior, and required by spec. |
|||||
►
Sign in to add a comment |
|||||
Comment 1 by mmenke@chromium.org
, Mar 14 2017Status: Untriaged (was: Unconfirmed)